[ASR pipeline] correct asr pipeline for seq2seq models (#15541)

src/transformers/pipelines/automatic_speech_recognition.py

@@ -265,10 +265,19 @@ class AutomaticSpeechRecognitionPipeline(ChunkPipeline):
             # it here.
             # Consume values so we can let extra information flow freely through
             # the pipeline (important for `partial` in microphone)
-            input_features = model_inputs.pop("input_features")
+            if "input_features" in model_inputs:
-            attention_mask = model_inputs.pop("attention_mask")
+                inputs = model_inputs.pop("input_features")
+            elif "input_values" in model_inputs:
+                inputs = model_inputs.pop("input_values")
+            else:
+                raise ValueError(
+                    "Seq2Seq speech recognition model requires either a "
+                    f"`input_features` or `input_values` key, but only has {model_inputs.keys()}"
+                )
+
+            attention_mask = model_inputs.pop("attention_mask", None)
             tokens = self.model.generate(
-                encoder_outputs=encoder(input_features=input_features, attention_mask=attention_mask),
+                encoder_outputs=encoder(inputs, attention_mask=attention_mask),
                 attention_mask=attention_mask,
             )
             out = {"tokens": tokens}

tests/test_pipelines_automatic_speech_recognition.py

@@ -107,6 +107,24 @@ class AutomaticSpeechRecognitionPipelineTests(unittest.TestCase, metaclass=Pipel
         output = speech_recognizer(waveform)
         self.assertEqual(output, {"text": "(Applaudissements)"})
 
+    @require_torch
+    def test_small_model_pt_seq2seq(self):
+        model_id = "hf-internal-testing/tiny-random-speech-encoder-decoder"
+        tokenizer = AutoTokenizer.from_pretrained(model_id)
+        feature_extractor = AutoFeatureExtractor.from_pretrained(model_id)
+
+        speech_recognizer = pipeline(
+            task="automatic-speech-recognition",
+            model=model_id,
+            tokenizer=tokenizer,
+            feature_extractor=feature_extractor,
+            framework="pt",
+        )
+
+        waveform = np.tile(np.arange(1000, dtype=np.float32), 34)
+        output = speech_recognizer(waveform)
+        self.assertEqual(output, {"text": "あл ش 湯 清 ه ܬ া लᆨしث ल eか u w 全 u"})
+
     @slow
     @require_torch
     @require_pyctcdecode